【摘要】 为解决护理机器人大小便监测采用单一传感器准确性不高而引发误报或漏报的问题,提出了一种BP神经网络和改进的D-S证据理论相结合的多传感器数据融合监测算法,并通过采集的湿度、温度和氨气浓度数据对融合模型进行仿真验证。首先利用BP神经网络对测量数据进行特征级监测,而后用BP网络的输出结果和训练误差获得D-S证据理论的基本信度分配,针对D-S证据理论无法解决证据之间的冲突问题,引入矛盾系数改进D-S证据理论实现决策级的融合监测。仿真结果表明:该方法降低了监测的不确定性,实现了异类信息的互补,提高了智能护理机器人排便监测的准确性和可靠性。更多还原

【Abstract】 In order to solve the problem of false alarms or omissions caused by the low accuracy of single sensors for defecation monitoring of nursing robots,a multi-sensor data fusion monitoring method based on BP neural network and improved D-S evidence theory is proposed,and the fusion model is verified by the collected humidity, temperature and ammonia concentration data. Firstly,the BP neural network is used to perform feature level monitoring on measurement data. Then the basic reliability distribution of the D-S evidence theory is calculated by using the output and the training errors of the BP network. In view of the fact that D-S evidence theory can not solve the conflict between evidence,and the contradiction coefficient is introduced to improve D-S evidence theory to achieve the fusion of decision-making level. The simulation results show that the method reduces the uncertainty of monitoring,realizes the complementarity of heterogeneous information,and improves the accuracy and reliability of the defecation monitoring of intelligent nursing robots.更多还原